The Problems In Wavelength Division Multiplexing Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

The Wavelength division multiplexing (WDM) technology is slowly increasing and improving in the last few years. Its existing system provides unusually great amount of bandwidth in a single fiber. In WDM (Wavelength Division Multiplexed) networks, there is a need to establish a proper routing and wavelength assignment procedure in order to achieve the most efficient communication. The problem taken into consideration is the RWA [5] problem wherein, for a given optical network and network connections, we have to assign the wavelength and select a suitable path in such a way that no two paths using the same wavelength should pass through the same link. WDM wide-area networks are very useful for increasing the bandwidth considerably. Using wavelength multiplexers we can an access node may transmit signals on different wavelength. However, the electronic switching and processing costs at the nodes can potentially be very high leading to severe performance bottlenecks and limiting the delivery of optical link bandwidth to the end users. This paper discusses a few RWA algorithms proposed in the past few years.


In WDM there is a need to maximize the network connections and minimize the blocking probability using a limited resource. WDM is a very promising technique that could be helpful in efficient data transmission in future large networks. Since, multiple data communications have to be achieved, we have to use various wavelength channels on the fiber [1].

This is critical and is generally achieved by the use of WDM technique as this could enhance the line capacity of the networks.


In WDM network there are three types of major constraints that are of prime concern.

Wavelength continuity constraint (WCC)

Distinct wavelength assignment constraint (DWAC)

No wavelength continuity constraint (NWCC).

In WCC same wavelength should be used on all links along the same path. In (DWAC) same wavelength cannot be used in two light paths on any fiber. In NWCC different wavelengths can be used along different links in the networks provided they have wavelength conversion capability. Eliminating the wavelength conversion in the network reduces the cost of setup significantly but it also has a negative effect on the efficiency of the network as more wavelengths may be required. But studies report that the efficiency loss is very minute in comparison to the cost effectiveness achieved. In a WDM optic network two access stations communicate through the use of light paths. A network that allows for the establishment light paths is often called an 'All Optical Network' [2]. The RWA problem is to assign proper route to the light path and to assign the wavelength [5]. The RWA problem can be studied in accordance to two heads i.e. static traffic and dynamic traffic[1,2].In case of static traffic a prior knowledge of the light path requests is there and thus routing can be done based on this knowledge of the traffic to be served by the network. In case of dynamic traffic, routing decisions and wavelength assignment will have to be made independent of all the paths that have already been assigned or will be assigned in future. The motive is to minimize the used wavelength and call blocking probability and increase the number of connections in order to make the whole system efficient. If a network is incapable of wavelength conversion and the wavelength will have to be chosen that is continuous along the nodes in the network. Now since the throughput is to be maximized and the blocking probability is to be minimized, it is a tightly coupled problem [1, 2].

This problem will now be studied under 2 heads: routing sub problem and wavelength assignment sub problem. Now let's look at each one at a time.


In this we select the most appropriate routes out of all the available routes. If multiple choices are available then an algorithm can be brought into play. The shortest path for each source destination pair is computed off-line in advance using standard shortest-path algorithms, e.g., Dijkstra's algorithm or Bellman-Ford algorithm [3, 4]. The disadvantage of this approach is that it does help in keep in mind the current state of the network. Thus, some links are underutilized and some are over-utilized [1].


This precisely means to assign the wavelength to the selected route. This step is specifically important as the proper implementation of the wavelength can minimize the use of wavelength convertors. Thus, it reduces a lot of cost involved. Whenever a call for transfer of data happens it goes to the controller which in turn has the details of the entire wavelength available and the routes to choose from. It is a very important task as it has a major effect on the performance. It is generally the First-fit algorithm that takes this decision [1, 3, 4].


A: Set of nodes in the network.

B: Set of links in the network.

C: Set of connections.

D: Set of wavelengths.

(NWAV): Total number of wavelengths numbered from 0 to N −1.

(ICONN): Total number of connection requests numbered from 0 to I −1.

(NODE): Total number of nodes in the network numbered from 0 to n −1.

(NLINKS): Total number of links along the route for sd connection.

S: indicates the source.

D: indicates the destination.

(SD [JCONN]: indicates jth connection.

(ROUTE [IJ]: represents the route for the connection when s = i and

d = j.

(WASIGN [IJ]): represents the wavelength assigned to the connection

When s = i and d = j.

[RECONN]: is the variable used to store the number of connections rejected.

acconn is the variable used to store the number of connections accepted.


Almost all the networks will employ shortest path strategy but alternate paths are used if the shortest path is not available. This might also happen in case of a trial to reduce the blocking probability. The best strategy is that the shortest path algorithm is used for routing and First Fit algorithm for wavelength [3]. In this method a try is made to establish with the first wavelength encountered. If unsuccessful then the second and so on till the last wavelength.

Now let us look at the proposed algorithms:


In this technique all the connection request is tried in the sequence using shortest path algorithm. Now this is repeated for all connection requests for all wavelengths. This method is again executed in a similar fashion but with an alternate path and not with the shortest path.


reconn ← I

for k ← 0 to N − 1

for j ← 0 to I − 1.

if connection for sd[jconn] not established earlier

then try to establish the connection for sd[jconn] on wave-length k for shortest path

else go to step 6.

if connection established in step 4

then reconn ← reconn - 1.

end loop for j .

end loop for k.

for k ← 0to N − 1.

for j ← 0to I − 1.

if connection for sd[jconn] not established earlier

then try to establish the connection for sd[jconn] on wave-length k for alternate path

else go to step 12.

if connection established for sd[jconn] in step 10

then reconn ← reconn - 1.

end loop for j .

end loop for k.

blocking probability ← reconn / I .



In this strategy the first connection request is tried by shortest path algorithm on all wavelengths .If unsuccessful then the same connection is tried on all different routes for all wavelengths. The same process is repeated for all processes.


reconn ← 0

for j ← 0to I − 1

for k ← 0 to N − 1

try to establish the connection for sd[jconn] on wavelength k for shortest path

if connection established in step 4 then go to step 12

end loop for k.

for k ← 0 to N − 1.

try to establish the connection for sd[jconn] on wavelength k for alternate shortest path

if connection established in step 8 then go to step 12

end loop for k

reconn ← reconn + 1

end loop for j

blocking probability ← reconn / I



In this method all the requests are tried on the first wavelength. And similarly all the rest of the connections are tried on same wavelengths for all routes.


reconn ← I

for k ← 0 to N − 1

for j = 0to I − 1

if connection for sd[jconn] not established with lower wave-length(s)

then try to establish the connection for sd[jconn] on wave- length k for shortest path

else go to step 6

if connection established in step 4

then reconn ← reconn - 1

end loop for j

for j ← 0 to I − 1

if connection for sd[jconn] not established earlier

then try to establish the connection for sd[jconn] on wave-length k for alternate shortest path

else go to step 10.

if connection established in step 8

then reconn ← reconn - 1

end loop for j .

end loop for k.

blocking probability ← reconn / I .



The first request is tried on the first wavelength using shortest path strategy. If unsuccessful the same request is tried on same wavelength using alternate routes. The same procedure is used for all wavelengths.


reconn = 0

for j = 0 to I − 1

for k = 0 to N − 1

try to establish the connection for sd[jconn] on wavelength k for shortest path

if connection established in step 4 then go to step 10

try to establish the connection for sd[jconn] on wavelength k for alternate path

if connection established in step 6 then go to step 10

end loop for k.


end loop for j

blocking probability = recon / I



Basically WDM can be viewed as a collection of optimization problems. We can solve these problems using an Evolutionary (GENETIC) algorithm. It's very helpful for solving the optimization problem. A special cost of function which frequency provides us fitness of a chromosome. M-point is very useful for the maintaining the verity of the solution space. The wavelength assign-to light paths in fittest individuals are performed using a special graph-coloring technique. The first can we Compared by obtaining the objective results. There are some algorithm like ARPANet, EON, UKNet and USFNet, all are part of the optical network. The results and the comparison show the effectiveness of algorithm.

Most of the source-destination pairs in a network assign single traffic. All optical channel of light path is very useful not only for circuit-switched traffic but also its extent over an area of multiple fiber links. The explanation of wavelength convertors, a light path is on the mind of the same wavelength through passes by the links. In the routing and wavelength assignment problem there is light path where we have to route and assign a wavelength to each of them and it can be developed according to an orderly plan like integer linear program like mixed-linear program there is assignment made less in size not only for node coloring problem in a graph but also NP complete. WA problem easily can solve these problems solve By Heuristics [6]. Suppose the problems are computationally very typical that time we can here for large problem [5]. RWA provides a wide range of optimization to solve various other optical network optimization problems. Traditional optimization methods useful for find the global optimum. Here we can say that the general ways are mesh network topology useful for minimizing the network cost is studied these are the use of the Evolutionary algorithms. Here a major difference between genetic algorithms and Genetic algorithms is that genetic algorithm gives better performance than simulated annealing.


Graph is collection of edge and vertices. Here vertices are denoted by V and edges are denoted by E. Here input parameter problem refer to here physical topology of an optical network. A set is made by a source and a destination.


There are two types of RWA problem first is wavelength continuity constraint these are perform a particular Request for a source to destination depends on the single light path and the another one Constraint is that The basic concept of Wavelength Conflict Constraint the two signals cannot be traverse in the particular fiber if wavelength is same[6].


Graph is made by vertices and edges. In the graph vertices is called nodes .Graph edges is called link. Here we are denoting Vertices and edges respectively V and E. These are represented by the network topology. The RWTA algorithm is very useful for established the light path [7].


FIXED ROUTING (FR): We want to find the shortest path between the source node and the destination node. That's why we use the shortest path routing algorithm such as Dijkstra's algorithm for routing purpose. This type of path that is called primary path and this type of approach is fixed shortest-path routing [3, 4].

ALTERNATE ROUTING (AR): The alternate way is to show that in the next shortest path between the source and destination, we cannot share links also with the primary shortest path.

DYNAMIC ROUTING (DR): Here we choose the path dynamically from the source to the destination node. Dynamic routing is depending on the network state these are responsible for all connections that are currently. It's all about depends the link loading, weight functions these all are associated with each link. The Least Resistance weight function decides on the path among primary and the alternate [6].


In the case of wavelength assignment all wavelengths are numbered and we use the First Fit algorithm. Now we can say that if when searching for available wavelength then we will consider the first before higher numbered wavelength. This scheme packs all of the in-use wavelengths towards the lower end of the wavelength space. It has a very cheap computation cost [3, 6].


First of all we will determine the wavelength than we allocate whatever the required time-slots. In the time-slot-algorithm there first fit algorithm will be used. It is similar to the first fit wavelength assignment algorithm. This algorithm searches the time-slots from lower-numbered slots to higher-numbered slot, and the first available time-slot is selected. All the transmission slots start of the frame. It has a cheapest computation cost [7].


The proposed reassignment algorithm for single-fiber TDM-WDM based on the optical backbone networks that support dynamic traffic.

The Link Status Matrix ΨL, K, F, is a three dimensional matrix, where L is the number of links in the network. Here K denotes the number of the wavelengths per link and F denotes the number of time-slots per wavelength. For example, the call to be established in the network, what are the links involved in the path, which wavelength is assigned in those links and which are the time-slots allocated in that particular wavelength[6,7].

The congestion level of time-slot t, Ct: The congestion level of any time-slot t, t = 1, 2 . . . F, is the number of wavelengths using the timeslot t. Maximum congestion can be Ct = K - F and minimum can be Ct = 0 which is called least congested timeslot.

Link Utilization (U): The link utilization for the network at any point of time is given by


To accommodate the continuous call blocking, we have to use MOLC reassignment algorithm, and first try and make the established calls from the minimum overlapping wavelength time length. The single wavelength with multiple wavelengths per link presented the pseudo code description of the algorithm. It's extended for multiple wavelengths per link if a new call gets blocked.

Step 1: Find the links required for the new call by using the routing schemes discussed before.

Step 2: From the Link Status Matrix ΨL, K, F, make a save-point at that with current time stamp say Ψ', determine how many of the required links are occupied by the already established calls in each timeslot. Store this information in the occupancy table.

Step 3: Sort the occupancy table in the ascending order, so that the timeslot, which has minimum overlap with, the required links of the new call is at the top of the table.

Step 4: Sort the timeslots according to the congestion level in the ascending order, such that the least congested timeslot is at the top of the congestion array.

Step 5: Try to reassign the already existing calls that are occupying the required links in the minimum overlapping timeslot (obtained from step 3) to the least congested timeslot.

If the calls could not be reassigned, try to move the call to the next least congested time slot.

If all the required links in the minimum overlapping timeslot are freed by reassignment, then establish the new call in the freed timeslot. Decrease n by 1.

If n is zero

then All the n requested slots for the new call are established in the network

Go to step 6.

Else Go to step 2.

If there is a next minimum overlapping timeslot available from the occupancy table

then Go to step 6.


Declare that the new call is blocked.

Change the current Link Status Matrix Ψ to Ψ'.

Step 6: Go back to the normal RWTA for the next call request arrival.


FR: This is part of the RWTA algorithm. It's used for fixed routing for the traffic demand. As in multi-rate (MR) traffic demand, the bandwidth demand of the session is uniformly distributed between 1 to FM with the requested number of timeslot varying up to a value. It should be at most F [6, 7].

FR-MOLC-AR-MOLC (FM-AM): It's very useful for removing the blocking due to continuity in the shortest path between in the source and the destination and is called FR-MOLC. If the call gets blocked in FR-MOLC these are useful for the alternate blocked call.

FR-SR: It just performs the basic operation of the routing wavelength and time slot reassignment algorithm. The basic purpose of this algorithm is that fixed routing for the single rate traffic demands that why we use this algorithm.

DR: The basic purpose of the dynamic routing is that, first we have to establish the light path for multiple traffic designs. Here we just making the light path connection because we want to design multiple traffic i.e. more than one traffic. After that, session gives the permission send the request for time slots a value. It should less than F.

DR-MOLC: The basic purpose of this algorithm is that removes the blocking due to the continuing constraints. For instance, if the new call is blocked, we will use the MOLC reassignment algorithm that will make a place for the blocked call in the network.


The dynamic routing and wave length can contain the different formations and their solving solutions algorithms to their optical networks. We can build large number of optical networks in the routing wavelength networks.

Light paths carry data at very high rates and stay for long time. It has been seen that when reassignment algorithm is used in the fixed and alternate routing, the performance is able to improve close to the single-rate traffic. The performance of the dynamic routing with the reassignment algorithm is also close to the single-rate traffic with fixed routing performance. Simulation results show that by with these techniques, it is possible to avoid and delay the time before which the first call block occurs, thus accommodating more calls in the network before network upgrading.